radeon/ttm/PCIe: Use dma_addr if TTM has set it.

If the TTM layer has used the DMA API to setup pages that are
TTM_PAGE_FLAG_DMA32 (look at patch titled: "ttm: Utilize the dma_addr_t
array for pages that are to in DMA32 pool."), lets use it
when programming the GART in the PCIe type cards.

This patch skips doing the pci_map_page (and pci_unmap_page) if
there is a DMA addresses passed in for that page. If the dma_address
is zero (or DMA_ERROR_CODE), then we continue on with our old
behaviour.

[v2: Fixed an indentation problem, added reviewed-by tag]
[v3: Added Acked-by Jerome]

Acked-by: Jerome Glisse <j.glisse@gmail.com>
Reviewed-by: Thomas Hellstrom <thomas@shipmail.org>
Signed-off-by: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com>
Tested-by: Ian Campbell <ian.campbell@citrix.com>

3 files changed, 33 insertions, 12 deletions

diff --git a/drivers/gpu/drm/radeon/radeon.h b/drivers/gpu/drm/radeon/radeon.h
index 73f600d39ad..c9bbab921e6 100644
--- a/drivers/gpu/drm/radeon/radeon.h
+++ b/drivers/gpu/drm/radeon/radeon.h
@@ -317,6 +317,7 @@ struct radeon_gart {
 	union radeon_gart_table		table;
 	struct page			**pages;
 	dma_addr_t			*pages_addr;
+	bool				*ttm_alloced;
 	bool				ready;
 };
 
@@ -329,7 +330,8 @@ void radeon_gart_fini(struct radeon_device *rdev);
 void radeon_gart_unbind(struct radeon_device *rdev, unsigned offset,
 			int pages);
 int radeon_gart_bind(struct radeon_device *rdev, unsigned offset,
-		     int pages, struct page **pagelist);
+		     int pages, struct page **pagelist,
+		     dma_addr_t *dma_addr);
 
 
 /*
diff --git a/drivers/gpu/drm/radeon/radeon_gart.c b/drivers/gpu/drm/radeon/radeon_gart.c
index e65b90317fa..5214bc29d9b 100644
--- a/drivers/gpu/drm/radeon/radeon_gart.c
+++ b/drivers/gpu/drm/radeon/radeon_gart.c
@@ -149,8 +149,9 @@ void radeon_gart_unbind(struct radeon_device *rdev, unsigned offset,
 	p = t / (PAGE_SIZE / RADEON_GPU_PAGE_SIZE);
 	for (i = 0; i < pages; i++, p++) {
 		if (rdev->gart.pages[p]) {
-			pci_unmap_page(rdev->pdev, rdev->gart.pages_addr[p],
-				       PAGE_SIZE, PCI_DMA_BIDIRECTIONAL);
+			if (!rdev->gart.ttm_alloced[p])
+				pci_unmap_page(rdev->pdev, rdev->gart.pages_addr[p],
+				       		PAGE_SIZE, PCI_DMA_BIDIRECTIONAL);
 			rdev->gart.pages[p] = NULL;
 			rdev->gart.pages_addr[p] = rdev->dummy_page.addr;
 			page_base = rdev->gart.pages_addr[p];
@@ -165,7 +166,7 @@ void radeon_gart_unbind(struct radeon_device *rdev, unsigned offset,
 }
 
 int radeon_gart_bind(struct radeon_device *rdev, unsigned offset,
-		     int pages, struct page **pagelist)
+		     int pages, struct page **pagelist, dma_addr_t *dma_addr)
 {
 	unsigned t;
 	unsigned p;
@@ -180,15 +181,22 @@ int radeon_gart_bind(struct radeon_device *rdev, unsigned offset,
 	p = t / (PAGE_SIZE / RADEON_GPU_PAGE_SIZE);
 
 	for (i = 0; i < pages; i++, p++) {
-		/* we need to support large memory configurations */
-		/* assume that unbind have already been call on the range */
-		rdev->gart.pages_addr[p] = pci_map_page(rdev->pdev, pagelist[i],
+		/* On TTM path, we only use the DMA API if TTM_PAGE_FLAG_DMA32
+		 * is requested. */
+		if (dma_addr[i] != DMA_ERROR_CODE) {
+			rdev->gart.ttm_alloced[p] = true;
+			rdev->gart.pages_addr[p] = dma_addr[i];
+		} else {
+			/* we need to support large memory configurations */
+			/* assume that unbind have already been call on the range */
+			rdev->gart.pages_addr[p] = pci_map_page(rdev->pdev, pagelist[i],
 							0, PAGE_SIZE,
 							PCI_DMA_BIDIRECTIONAL);
-		if (pci_dma_mapping_error(rdev->pdev, rdev->gart.pages_addr[p])) {
-			/* FIXME: failed to map page (return -ENOMEM?) */
-			radeon_gart_unbind(rdev, offset, pages);
-			return -ENOMEM;
+			if (pci_dma_mapping_error(rdev->pdev, rdev->gart.pages_addr[p])) {
+				/* FIXME: failed to map page (return -ENOMEM?) */
+				radeon_gart_unbind(rdev, offset, pages);
+				return -ENOMEM;
+			}
 		}
 		rdev->gart.pages[p] = pagelist[i];
 		page_base = rdev->gart.pages_addr[p];
@@ -251,6 +259,12 @@ int radeon_gart_init(struct radeon_device *rdev)
 		radeon_gart_fini(rdev);
 		return -ENOMEM;
 	}
+	rdev->gart.ttm_alloced = kzalloc(sizeof(bool) *
+					 rdev->gart.num_cpu_pages, GFP_KERNEL);
+	if (rdev->gart.ttm_alloced == NULL) {
+		radeon_gart_fini(rdev);
+		return -ENOMEM;
+	}
 	/* set GART entry to point to the dummy page by default */
 	for (i = 0; i < rdev->gart.num_cpu_pages; i++) {
 		rdev->gart.pages_addr[i] = rdev->dummy_page.addr;
@@ -267,6 +281,8 @@ void radeon_gart_fini(struct radeon_device *rdev)
 	rdev->gart.ready = false;
 	kfree(rdev->gart.pages);
 	kfree(rdev->gart.pages_addr);
+	kfree(rdev->gart.ttm_alloced);
 	rdev->gart.pages = NULL;
 	rdev->gart.pages_addr = NULL;
+	rdev->gart.ttm_alloced = NULL;
 }
diff --git a/drivers/gpu/drm/radeon/radeon_ttm.c b/drivers/gpu/drm/radeon/radeon_ttm.c
index 6f156e9d3f3..ca045058e49 100644
--- a/drivers/gpu/drm/radeon/radeon_ttm.c
+++ b/drivers/gpu/drm/radeon/radeon_ttm.c
@@ -647,6 +647,7 @@ struct radeon_ttm_backend {
 	unsigned long			num_pages;
 	struct page			**pages;
 	struct page			*dummy_read_page;
+	dma_addr_t			*dma_addrs;
 	bool				populated;
 	bool				bound;
 	unsigned			offset;
@@ -662,6 +663,7 @@ static int radeon_ttm_backend_populate(struct ttm_backend *backend,
 
 	gtt = container_of(backend, struct radeon_ttm_backend, backend);
 	gtt->pages = pages;
+	gtt->dma_addrs = dma_addrs;
 	gtt->num_pages = num_pages;
 	gtt->dummy_read_page = dummy_read_page;
 	gtt->populated = true;
@@ -674,6 +676,7 @@ static void radeon_ttm_backend_clear(struct ttm_backend *backend)
 
 	gtt = container_of(backend, struct radeon_ttm_backend, backend);
 	gtt->pages = NULL;
+	gtt->dma_addrs = NULL;
 	gtt->num_pages = 0;
 	gtt->dummy_read_page = NULL;
 	gtt->populated = false;
@@ -694,7 +697,7 @@ static int radeon_ttm_backend_bind(struct ttm_backend *backend,
 		     gtt->num_pages, bo_mem, backend);
 	}
 	r = radeon_gart_bind(gtt->rdev, gtt->offset,
-			     gtt->num_pages, gtt->pages);
+			     gtt->num_pages, gtt->pages, gtt->dma_addrs);
 	if (r) {
 		DRM_ERROR("failed to bind %lu pages at 0x%08X\n",
 			  gtt->num_pages, gtt->offset);